Determining the Choice of Entry Mode of Multinationals : My 6 Million Regressions

Motivated by the findings of Ramada-Sarasola (2009) and the lack of robustness of the previous literature's results on foreign entry mode choice to model specification in this paper I perform an Extreme Bounds Analysis to determine which of almost 60 explanatory variables used in the literature are robust to different model specifications. I do so by following the methodology introduced in Sala-i-Martin (1997b) in a multinomial logit framework, based on 640 entries into foreign countries done by the largest 22 financial MNCs in the last 15 years. I suggest additional hypothesis to capture host-country level determinants and I improve the operationalization of industry level variables in a multi home-country and host-country setting. Amongst other results I find that an MNC's size and its international experience increase the likelihood of greenfields as opposed to M&A or any type of entry mode involving a partner, and that more cultural distance between home and host country, a better developed local financial sector (or local credit market), a more regulated environment for obtaining licenses and more macroeconomic sustainability increase the chances of GI, while a worse local infrastructure, higher ITC costs and more difficulties in registering property and employing workers decrease its odds